spiceai

Why we included this project

Building AI features that answer questions from your own data usually means wiring several pieces of infrastructure together yourself. Spice removes that work: it is a single Rust runtime you run as a sidecar next to your app, serving SQL, search, and LLM-inference requests over the operational data you already have. Rather than copying data into a separate store, it federates queries across your live sources, so one portable binary covers both accelerated analytics and local model inference. That suits agents and applications that need low-latency answers against production data, and it keeps a retrieval-style prototype far simpler to stand up.